Re: [openstack-dev] Recommended ways to find compute node's bandwidth (physical NIC)
Thanks Jay for pointing this out. Adnan, Yeah - you could use PcP - if that module gets integrated with Nova. For your benefit, here are some of the metrics that PcP can provide related to the physical NIC: pminfo | grep network.interface <== Listing the available metrics. network.interface.collisions network.interface.mtu network.interface.speed network.interface.baudrate network.interface.duplex network.interface.up network.interface.running network.interface.inet_addr network.interface.ipv6_addr network.interface.ipv6_scope network.interface.hw_addr network.interface.in.bytes network.interface.in.packets network.interface.in.errors network.interface.in.drops network.interface.in.fifo network.interface.in.frame network.interface.in.compressed network.interface.in.mcasts network.interface.out.bytes network.interface.out.packets network.interface.out.errors network.interface.out.drops network.interface.out.fifo network.interface.out.carrier network.interface.out.compressed network.interface.total.bytes network.interface.total.packets network.interface.total.errors network.interface.total.drops network.interface.total.mcasts # pmval network.interface.out.bytes <== A sample metric value. metric:network.interface.out.bytes host: semantics: cumulative counter (converting to rate) units: byte (converting to byte / sec) samples: all wlp3s0lo virbr0 wc0 em1 virbr0-nic 0.0 0.0 0.0 0.0 0.0 0.0 85.96 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1588. 4031. 0.0 787.6 0.0 0.0 492.7 0.0 0.0 155.9 0.0 0.0 As you might observe, there's a plenty for you to draw inferences. Hope this helps! Cheers, Sudipto On Monday 20 June 2016 09:07 AM, Jay Pipes wrote: On 06/16/2016 04:16 PM, KHAN, RAO ADNAN wrote: I am writing a nova filter that will check for the compute node (max, avg) bandwidth, before instantiating an instance. What are some of the recommended tools that can provide this info in real time? Does any openstack component hold this info already? You may want to chat with Sudipta Biswas (cc'd). The PCP module (yes, it's an awkward name for a performance profiling module...) does some of what you're looking for and Sudipta has been attempting to integrate it with Nova for this express purpose. Best, -jay __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Higgins][Zun] Project roadmap
On Monday 13 June 2016 06:57 PM, Flavio Percoco wrote: On 12/06/16 22:10 +, Hongbin Lu wrote: Hi team, During the team meetings these weeks, we collaborated the initial project roadmap. I summarized it as below. Please review. * Implement a common container abstraction for different container runtimes. The initial implementation will focus on supporting basic container operations (i.e. CRUD). What COE's are being considered for the first implementation? Just docker and kubernetes? * Focus on non-nested containers use cases (running containers on physical hosts), and revisit nested containers use cases (running containers on VMs) later. * Provide two set of APIs to access containers: The Nova APIs and the Zun-native APIs. In particular, the Zun-native APIs will expose full container capabilities, and Nova APIs will expose capabilities that are shared between containers and VMs. - Is the nova side going to be implemented in the form of a Nova driver (like ironic's?)? What do you mean by APIs here? - What operations are we expecting this to support (just CRUD operations on containers?)? I can see this driver being useful for specialized services like Trove but I'm curious/concerned about how this will be used by end users (assuming that's the goal). * Leverage Neutron (via Kuryr) for container networking. * Leverage Cinder for container data volume. * Leverage Glance for storing container images. If necessary, contribute to Glance for missing features (i.e. support layer of container images). Are you aware of https://review.openstack.org/#/c/249282/ ? This support is very minimalistic in nature, since it doesn't do anything beyond just storing a docker FS tar ball. I think it was felt that, further support for docker FS was needed. While there were suggestions of private docker registry, having something in band (w.r.t openstack) maybe desirable. * Support enforcing multi-tenancy by doing the following: ** Add configurable options for scheduler to enforce neighboring containers belonging to the same tenant. ** Support hypervisor-based container runtimes. The following topics have been discussed, but the team cannot reach consensus on including them into the short-term project scope. We skipped them for now and might revisit them later. * Support proxying API calls to COEs. Any link to what this proxy will do and what service it'll talk to? I'd generally advice against having proxy calls in services. We've just done work in Nova to deprecate the Nova Image proxy. * Advanced container operations (i.e. keep container alive, load balancer setup, rolling upgrade). * Nested containers use cases (i.e. provision container hosts). * Container composition (i.e. support docker-compose like DSL). NOTE: I might forgot and mis-understood something. Please feel free to point out if anything is wrong or missing. It sounds you've got more than enough to work on for now, I think it's fine to table these topics for now. just my $0.02 Flavio __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [dragonflow] Low OVS version for Ubuntu
On Thursday 17 September 2015 12:22 PM, Li Ma wrote: Hi all, I tried to run devstack to deploy dragonflow, but I failed with lower OVS version. I used Ubuntu 14.10 server, but the official package of OVS is 2.1.3 which is much lower than the required version 2.3.1+? So, can anyone provide a Ubuntu repository that contains the correct OVS packages? Why don't you just build the OVS you want from here: http://openvswitch.org/download/ Thanks, __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [nova] Regarding NUMA Topology filtering logic.
Hi, Currently the numa_topology filter code in openstack is going by a decision of filtering out NUMA nodes based on the length of the cpusets on the NUMA node of a host[1]. For example: if a VM with 8 VCPUs is requested, we seem to be doing len(cputset_on_the_numa_node) should be greater than or equal to 8. IMHO, the logic can be further improved if we start taking the threads and cores into consideration instead of directly going by the cpuset length of the NUMA node. This thought is derived from an architecture like ppc where each core can have 8 threads. However in this case, libvirt reports only 1 thread out of the 8 (called the primary thread). The host scheduling of the guests happen at the core level(as only primary thread is online). The kvm scheduler exploits as many threads of the core as needed by guest. Consider an example for the ppc architecture. In a given NUMA node 0 - with 40 threads - the following cpusets would be reported by libvirt: 0, 8, 16, 24, 32. The length of the cpusets would suggest that only 5 pcpus are available for pinning, however we could potentially have 40 threads available (cores * threads). This way we could at least solve the problem that arises if we just take the length of the cpusets into consideration. Thoughts? [1] https://github.com/openstack/nova/blob/master/nova/virt/hardware.py#L772 Thanks, Sudipto __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Neutron] Regarding neutron bug # 1432582
Thanks, I have got a patchset out for review. I have removed the exception that was being thrown back to the agent and have reduced the fix to just logging a meaningful message in the neutron server logs. Appreciate your comments on the same. Thanks, Sudipto On Monday 13 April 2015 11:56 AM, Kevin Benton wrote: I would like to see some form of this merged at least as an error message. If a server has a bad CMOS battery and suffers a power outage, it's clock could easily be several years behind. In that scenario, the NTP daemon could refuse to sync due to a sanity check. On Wed, Apr 8, 2015 at 10:46 AM, Sudipto Biswas sbisw...@linux.vnet.ibm.com mailto:sbisw...@linux.vnet.ibm.com wrote: Hi Guys, I'd really appreciate your feedback on this. Thanks, Sudipto On Monday 30 March 2015 12:11 PM, Sudipto Biswas wrote: Someone from my team had installed the OS on baremetal with a wrong 'date' When this node was added to the Openstack controller, the logs from the neutron-agent on the compute node showed - AMQP connected. But the neutron agent-list command would not list this agent at all. I could figure out the problem when the neutron-server debug logs were enabled and it vaguely pointed at the rejection of AMQP connections due to a timestamp miss match. The neutron-server was treating these requests as stale due to the timestamp of the node being behind the neutron-server. However, there's no good way to detect this if the agent runs on a node which is ahead of time. I recently raised a bug here: https://bugs.launchpad.net/neutron/+bug/1432582 And tried to resolve this with the review: https://review.openstack.org/#/c/165539/ It went through quite a few +2s after 15 odd patch sets but we still are not in common ground w.r.t addressing this situation. My fix tries to log better and throw up an exception to the neutron agent on FIRST time boot of the agent for better detection of the problem. I would like to get your thoughts on this fix. Whether this seems legit to have the fix per the patch OR could you suggest a approach to tackle this OR suggest just abandoning the change. __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev -- Kevin Benton __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
Re: [openstack-dev] [Neutron] Regarding neutron bug # 1432582
Hi Guys, I'd really appreciate your feedback on this. Thanks, Sudipto On Monday 30 March 2015 12:11 PM, Sudipto Biswas wrote: Someone from my team had installed the OS on baremetal with a wrong 'date' When this node was added to the Openstack controller, the logs from the neutron-agent on the compute node showed - AMQP connected. But the neutron agent-list command would not list this agent at all. I could figure out the problem when the neutron-server debug logs were enabled and it vaguely pointed at the rejection of AMQP connections due to a timestamp miss match. The neutron-server was treating these requests as stale due to the timestamp of the node being behind the neutron-server. However, there's no good way to detect this if the agent runs on a node which is ahead of time. I recently raised a bug here: https://bugs.launchpad.net/neutron/+bug/1432582 And tried to resolve this with the review: https://review.openstack.org/#/c/165539/ It went through quite a few +2s after 15 odd patch sets but we still are not in common ground w.r.t addressing this situation. My fix tries to log better and throw up an exception to the neutron agent on FIRST time boot of the agent for better detection of the problem. I would like to get your thoughts on this fix. Whether this seems legit to have the fix per the patch OR could you suggest a approach to tackle this OR suggest just abandoning the change. __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
[openstack-dev] [Neutron] Regarding neutron bug # 1432582
Someone from my team had installed the OS on baremetal with a wrong 'date' When this node was added to the Openstack controller, the logs from the neutron-agent on the compute node showed - AMQP connected. But the neutron agent-list command would not list this agent at all. I could figure out the problem when the neutron-server debug logs were enabled and it vaguely pointed at the rejection of AMQP connections due to a timestamp miss match. The neutron-server was treating these requests as stale due to the timestamp of the node being behind the neutron-server. However, there's no good way to detect this if the agent runs on a node which is ahead of time. I recently raised a bug here: https://bugs.launchpad.net/neutron/+bug/1432582 And tried to resolve this with the review: https://review.openstack.org/#/c/165539/ It went through quite a few +2s after 15 odd patch sets but we still are not in common ground w.r.t addressing this situation. My fix tries to log better and throw up an exception to the neutron agent on FIRST time boot of the agent for better detection of the problem. I would like to get your thoughts on this fix. Whether this seems legit to have the fix per the patch OR could you suggest a approach to tackle this OR suggest just abandoning the change. __ OpenStack Development Mailing List (not for usage questions) Unsubscribe: openstack-dev-requ...@lists.openstack.org?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev